Antigenic activity of dense bodies from human cytomegalovirus-infected cells in the complement fixation test.

I would like to describe briefly some preliminary serologic data that we have generated recently which may have some relevance to the cytomegalovirus (CMV) vaccine question. The CMV purification scheme used at the Community Blood Center starts with clarified, extracellular AD169 which is concentrated by ultrafiltration and then subjected to density gradient centrifugation in 20 to 60 percent sucrose gradients for 1 hr at 31,000 rpm in the SW50.1 rotor. Three light-scattering bqnds can be identified, the bottom most of which, at 45% sucrose, is comprised mainly of dense bodies. Dense bodies are electron-dense cytoplasmic material which has been described in cells infected in vitro with a number of herpes viruses. They have been seen in human CMV (1), rat CMV (2), murine CMV (3), varicella-zoster virus (4), Lucke virus of frogs (5), and Marek's disease virus (6). Although lysosomes are present in CMV-infected cells, they are morphologically distinct from dense bodies, a fact emphasized by Craighead, et al. (7) in a very careful study of dense bodies in cells infected with human CMV. They found no evidence that dense bodies contain lipids or carbohydrates and thus concluded that they are comprised largely of proteins. No comments were made as to their possible nucleic acid content. An interesting observation made by Craighead and co-workers was that by immune electron microscopy they could identify at least one cross-reactive antigenic determinant on enveloped CMV virions and dense bodies. This observation prompted us to use the material in the lowest sucrose band, that is, the one composed primarily of dense bodies, as a complement-fixation (CF) antigen to see if we could identify antibodies in human sera that would react with dense bodies. Table 1 summarizes the results of this experiment. Patient C is 6-months convalescent from a documented case ofCMV mononucleosis whose serum has consistently given a titer of 1:64 in the Laboratory Branch Complement Fixation (LBCF) test against a "standard" AD169 CF antigen, i.e., an alkaline glycine buffer extract of sonicated, infected human fibroblasts. When dense bodies were used as the CF antigen, this serum had a titer of 1:32. Patient H underwent a hysterectomy and was followed prospectively for evidence of transfusion-associated CMV infection. Her pretransfusion CMV CF titer was 1:64 and rose to 1:256 13 weeks after transfusion, at which time she was excreting CMV in her urine. Testing of these sera using dense bodies as the CF antigen yielded titers of 1:16 and 1:64, respectively. Patient M was referred to us with heterophile negative mononucleosis which was subsequently diagnosed as due to CMV. The virus was initially isolated from her blood and urine. A urine specimen taken 3 weeks